Human reserve pluripotent mesenchymal stem cells are present in the connective tissues of skeletal muscle and dermis derived from fetal, adult, and geriatric donors

Authors


Abstract

This study details the profile of 13 cell surface cluster differentiation markers on human reserve stem cells derived from connective tissues. Stem cells were isolated from the connective tissues of dermis and skeletal muscle derived from fetal, mature, and geriatric humans. An insulin/dexamethasone phenotypic bioassay was used to determine the identity of the stem cells from each population. All populations contained lineage-committed myogenic, adipogenic, chondrogenic, and osteogenic progenitor stem cells as well as lineage-uncommitted pluripotent stem cells capable of forming muscle, adipocytes, cartilage, bone, fibroblasts, and endothelial cells. Flow cytometric analysis of adult stem cell populations revealed positive staining for CD34 and CD90 and negative staining for CD3, CD4, CD8, CD11c, CD33, CD36, CD38, CD45, CD117, Glycophorin-A, and HLA DR-II. Anat Rec 264:51–62, 2001. © 2001 Wiley-Liss, Inc.

Reserve stem cells have been reported to be present in postnatal animals. Stem cells capable of reconstituting the tissues in which they reside include myosatellite cells in skeletal muscle (Grounds, 1999), chondroblasts in the perichondrium surrounding cartilage (Yotsuyanagi et al., 1999), oval cells in the liver (Gordon et al., 2000), pre-ductal stem cells in the pancreas (Bonner-Wier et al., 2000; Ramiya et al., 2000), basal stem cells of epithelial tissues (Slack, 2000), and neuronal stem cells in the brain (Gage, 2000). By contrast, other studies have shown that stem cells residing in one tissue can differentiate into cell types of another tissue. For example, reserve stem cells derived from bone marrow can differentiate into neurons and glia (Eglitis and Mezey, 1997; Kopen et al., 1999), restore dystrophin expression in skeletal muscle (Gussoni et al., 1999), serve as a source of hepatic oval cells (Petersen et al., 1999), provide cells for neovascularization (Asahara et al., 1999; Kalka et al., 2000), and restore cartilage, bone, and fat (Pittenger et al., 1999; Prokop, 1997; Yoo et al., 1998). Conversely, stem cells derived from neural tissues (Bjornson et al., 1999) and muscle tissues (Jackson et al., 1999) can differentiate into blood. Furthermore, reserve stem cells derived from dermis can differentiate into skeletal muscle, fat, cartilage and bone (Young et al., 1999). These studies suggest that the fates of some of these reserve stem cells are not restricted by their location within a particular tissue or organ.

Our current research is focused upon characterizing the cell surface antigens and capacity for differentiation of human reserve stem cells derived from the connective tissues of skeletal muscle and dermis. We have previously characterized the capacity for differentiation of putative stem cells using an insulin/dexamethasone bioassay (Lucas et al., 1993, 1995; Pate et al., 1993; Rogers et al., 1995; Warejcka et al., 1996; Young, 2000; Young et al., 1991, 1992a,1992b, 1993, 1995, 1998a,1998b, 1999). Numerous investigators have used antibodies to cluster of differentiation (CD) markers to characterize and isolate hematopoietic cells based on the profiles of their cell surface antigens (review, Kishimoto et al., 1997). CD markers have also been shown to be present on nonhematopoietic cells, including differentiated phenotypes, tumorigenic cells, and reserve stem cells. For example, CD34 has been reported to be present on vascular endothelial cells (Lin et al., 1995; Suda et al., 2000), thyroid interfollicular fibroblasts (Yamazaki and Eyden, 1997), tumorigenic cells (Ohtani et al., 1998), and marrow stromal cells (Waller et al., 1995). CD44 has been reported to be present on periosteal stromal cells (Ghilzon et al., 1999) and marrow mesenchymal progenitor cells (Conget and Minguell, 1999). CD51 and CD56 have been reported to be present on the endosteal cells of bone marrow (Sillaber et al., 1999). CD105 has been reported to be present on a tripotent adipogenic-chondrogenic-osteogenic progenitor cell from the stroma of bone marrow (Barry et al., 1999; Pittenger et al., 1999). CD10, CD13, CD56, and MHC Class-I have been identified on human reserve stem cells (Young et al., 1999).

The experiments reported in this study involve characterizing the capacity for differentiation of human reserve stem cells derived from the connective tissues of skeletal muscle and dermis of human fetal, adult, and geriatric donors. These reserve stem cells have been analyzed for the presence of cell surface markers for CD3, CD4, CD8, CD11c, CD33, CD34, CD36, CD38, CD45, CD90, CD117, Glycophorin-A, and HLA DR-II.

MATERIALS AND METHODS

Human Reserve Stem Cells Derived From Connective Tissues

Six populations of reserve stem cells derived from the connective tissues of skeletal muscle and dermis were used in these experiments. These populations were derived from fetal (female and male), mature (two female) and geriatric (female and male) donors. The protocols for harvesting human tissues have been approved by the Institutional Review Board at the Medical Center of Central Georgia, Macon, GA.

Cells derived from the connective tissues of skeletal muscle and dermis obtained from fetal and mature donors were purchased from Clonetics (San Diego, CA) and processed by differential plating and cryopreservation, yielding putative stem cells (Young et al., 1991, 1992a, 1999). In brief, the cells were transferred to plating medium-A (PM-A). PM-A consisted of 89% (v/v) Eagle's minimal essential medium with Earle's salts (EMEM, GIBCO BRL, Life Technologies, Grand Island, NY), 10% (v/v) pre-selected horse serum (lot numbers 17F-0218 [HS7] or 49F-0082 [HS4], Sigma Chemical Co., St. Louis, MO), and 1% (v/v) penicillin/streptomycin (10,000 U/ml penicillin and 10,000 mg/ml streptomycin, GIBCO), pH 7.4. Cells were incubated in a 95% air/5% CO2 humidified environment. After expansion, cells were released with 0.05% (w/v) trypsin (DIFCO, Becton-Dickinson Labware, Franklin Lakes, NJ) in Ca+2,Mg+2-free Dulbecco's phosphate buffered saline (GIBCO) containing 0.0744% (w/v) ethylenediamine tetraacetic acid (EDTA, Sigma), centrifuged at 100 × g for 20 min, and the supernatant aspirated. The cell pellet was resuspended in PM-A and the cell suspension cryopreserved by slow freezing for storage at −70 to −80°C in PM-A containing 7.5% (v/v) dimethyl sulfoxide (DMSO, Morton Thiokol, Danvers, MA) (Young et al., 1991). These cell lines were designated “CF-SkM1” (fetal female skeletal muscle connective tissue, CC-2561, lot number 14722, Clonetics), “CM-SkM1” (fetal male skeletal muscle connective tissue, CC-0231, lot number 6F0604, Clonetics), “NHDF1” (25-year-old female dermis, CC-0252, lot number 6F0600, Clonetics), and “NHDF2” (36-year-old female dermis, CC-0252, lot number 16280, Clonetics).

Geriatric cells were isolated from the endomysial, perimysial and epimysial connective tissue compartments associated with skeletal muscle specimens obtained from a 77-year-old female patient and a 67-year-old male patient following standard protocols for the isolation of mesenchymal stem cells (Lucas et al., 1995; Young et al., 1999). In brief, cells were liberated from the connective tissue compartment of skeletal muscle with collagenase (CLS-I, Worthington Biochemical Corp., Freehold, NJ) and dispase (Collaborative Biomedical Products, Bedford, MA). Single cell suspensions were obtained by sequential filtration through 90-μm and 20-μm Nitex (Tetco Inc., Elmsford, NY). Cells were seeded at 105 cells/1% (w/v) gelatin-coated (EM Sciences, Gibbstown, NJ) T-75 flasks (Falcon, Becton-Dickinson Labware) in PM-A and allowed to expand and differentiate before cryopreservation. Cells were incubated as before. After expansion, cells were released with trypsin, sieved as above to separate mononucleated cells from differentiated phenotypes (i.e., multinucleated myotubes, adipocyte colonies, cartilage nodules, bone nodules), and cryopreserved at −70 to −80°C in PM-A containing 7.5% (v/v) DMSO. Using the procedures outlined above, each subsequent cryopreservation step effectively removed more than 98% of contaminating fibroblasts and differentiated phenotypes from the stem cell preparation (Young et al., 1991). These cells were designated as “PAL2” and “PAL3”, respectively.

After initial harvest, further expansion of the cell lines through 30 cell doublings and more than 70 cell doublings was accomplished by repeated propagation and cryopreservation utilizing 1% gelatin coated flasks with plating medium-B (PM-B) (Young, 2000; Young et al., 1991, 1993, 1998b). PM-B consisted of 89% (v/v) Opti-MEM based medium (22600-050, GIBCO) containing 0.01 mM β-mercapto-ethanol (Sigma), 10% (v/v) horse serum (HS3, lot number 3M0338, BioWhittaker, Walkersville, MD), and 1% (v/v) antibiotic-antimycotic solution (10,000 U/ml penicillin, 10,000 μg/ml streptomycin, 25 μg/ml Amphotericin-B, GIBCO), pH 7.4. Cells were then propagated to 30 cell doublings and beyond, released with trypsin, and aliquoted for insulin/dexamethasone analyses and flow cytometric analyses. At each harvest, the cells were counted and the number of cells compared with the original plating densities. The number of cell doublings per passage was determined (Young, 2000; Young et al., 1991, 1993, 1995). Programmed senescence has been shown to occur at approximately 50–70 cell doublings (Hayflick's limit; Hayflick, 1965). The number of cell doublings was chosen so that one population would consist of cells at 30 cell doublings (less than Hayflick's limit). This population included both progenitor cells and pluripotent stem cells. Another population was chosen so that it would consist of cells beyond 70 cell doublings (at more than Hayflick's limit). As progenitor cells die around Hayflick's limit, the second population should be limited to pluripotent stem cells.

Insulin/Dexamethasone Bioassay

Cells were examined using an insulin/dexamethasone bioassay to determine the existence of progenitor or pluripotent stem cells within the populations examined (Young, 2000; Young et al., 1998a,1998b). In this bioassay, insulin accelerates the phenotypic expression of progenitor stem cells but has no effect on the induction of phenotypic expression in pluripotent stem cells. By contrast, dexamethasone induces lineage-commitment and expression in pluripotent stem cells, but does not alter phenotypic expression in progenitor stem cells (Young, 2000; Young et al., 1993, 1998a,1998b). Therefore, if progenitor cells alone are present in the culture there will be no difference in either the quality or quantity of phenotypes expressed upon treatment with insulin. Similar cultures will exhibit the expression of multiple phenotypes upon treatment with dexamethasone. If the culture is mixed, containing both progenitor and pluripotent cells, there will be a greater quality or quantity of phenotypes expressed upon treatment with dexamethasone compared with treatment with insulin. Thus, comparing the effects of treatment with insulin and dexamethasone permits the identification of specific types of progenitor and pluripotent stem cells within an unknown population of cells (Young, 2000; Young et al., 1998b, 1999).

Cells were processed as described previously (Young et al., 1999). In brief, aliquots of CM-SkM, CF-SkM, NHDF1, NHDF2, PAL#2, and PAL#3 cells were thawed and plated individually at 10,000 cells/well in 1% gelatin-coated 24-well plates (Corning, Corning, NY) utilizing PM-B. After 24 hr PM-B was removed and replaced with either control medium, insulin testing medium, or dexamethasone testing medium. Control medium consisted of 98% (v/v) Opti-MEM containing 0.01 mM β-mercapto-ethanol, 1% (v/v) HS3, and 1% antibiotic-antimycotic solution, pH 7.4. Insulin testing medium consisted of control medium containing 2 μg/ml insulin (Sigma), pH 7.4. Dexamethasone testing medium was composed of 98% Opti-MEM, 0.01 mM β-mercaptoethanol, 1% serum (HS3 or HS9, 90H-0701, Sigma) or FBS (fetal bovine serum, lot number 3000L, Atlanta Biologicals, Norcross, GA) and 1% antibiotic-antimycotic solution, pH 7.4. This solution was made 10−10, 10−9, 10−8, 10−7, or 10−6 M with respect to dexamethasone (Sigma) (Young, 2000; Young et al., 1995, 1998, 1999). Media were changed three times per week for 6 to 8 weeks. Cultures were viewed twice per week for changes in phenotypic expression and photographed.

The phenotypic expression of the cells was assayed morphologically. The morphological changes observed were identical to those previously noted in avian, mouse, rat, rabbit, and human progenitor and pluripotent mesenchymal stem cells after incubation with insulin or dexamethasone (Pate et al., 1993; Rogers et al., 1995; Young, 2000; Young et al., 1993, 1995, 1998a,1998b). Insulin induces a 1–5% expression in lineage-committed progenitor cells (Table 1) and dexamethasone induces a 6–10% expression in lineage-uncommitted cells (Tables 1 and 2). Phenotypes were expressed at all concentrations of dexamethasone from 10−10 to 10−6 M. Maximal expression of a particular phenotype, however, differed with respect to both dexamethasone concentration and time in culture.

Table 1. Induction of the expression of different mesodermal morphological phenotypes by insulin and dexamethasone in human stem cells at 30 cell doublingsa
 Insulin (2 μg/ml)Dexamethasone (10−10 to 10−6 M)
CF-SkM1CM-SkM1NHDF1NHDF2PAL2PAL3CF-SkM1CM-SkM1NHDF1NHDF2PAL2PAL3
  • a

    Cells, stem cells treated at 30 cell doublings after harvest.

  • b

    Multi-nuc, multinucleated linear and branched skeletal muscle myotubes; F5D, myogenin; MF-20, sarcomeric myosin; MY-32, fast skeletal muscle myosin; ALD58, myosin heavy chain; A4.74, myosin fast chain.

  • c

    +, Approximately 1–5% of culture expressing each particular designated phenotype, with approximately 20% of culture exhibiting all four phenotypes, i.e., skeletal muscle myotubes, adipocytes, cartilage nodules, and bone nodules.

  • d

    ++, Approximately 6–10% of culture expressing each particular designated phenotype, with ≤70% of culture exhibiting all eight phenotypes, i.e., skeletal myogenic, smooth myogenic, cardiogenic, adipogenic, chondrogenic, osteogenic, fibrogenic, and endotheliogenic.

  • e

    −, absence of phenotype or staining.

  • f

    PMwIF, large polygonal-shaped mononucleated cell containing intracellular filaments; IA4, smooth muscle α-actin.

  • g

    Binucleated, binucleated cells; IA4 + MF-20, colabeling of smooth muscle α-actin and sarcomeric myosin.

  • h

    PMwRICV, polygonal-shaped mononucleated cells with multiple refractile intracellular vesicles containing saturated neutral lipids; SBB, Sudan Black-B; ORO, Oil Red-O.

  • i

    Rd + PCMH, aggregations of rounded cells with refractile pericellular matrix halos; C11C1, collagen pro type-II; HC-II, Collagen type-II; D1-9, type-IX collagen; AB1.0, Alcian Blue at pH 1.0 (chondroitin sulfate and keratan sulfate glycosaminoglycans); PF/AIcBI, Perfix/Alcec Blue (sulfated glycosaminoglycans).

  • j

    3D-Matrix, three-dimensional extracellular matrices overlying cellular aggregations; WV1D1, bone sialoprotein-II; MP111, osteopontine; vK, von Kossa (calcium phosphate).

  • k

    SpM + PM, spindle-shaped and polygonal-shaped mononucleated cells; 1B10, human fibroblast surface protein.

  • l

    Cobblestone, individual or sheets of cobblestone-shaped or polygonal-shaped mononucleated cells; P1H12, human-specific endothelial cell surface marker; P2B1, peripheral endothelial cell adhesion molecule (PECAM); P8B1, vascular cell adhesion molecule (VCAM); P2H3, E-selectin.

Skeletal muscle
 Multi-nucb+c+++++++d++++++++++
 F5De++++++++++++
 MF-20++++++++++++++++++
 MY-32++++++++++++++++++
 ALD58++++++++++++++++++
 A4.74++++++++++++++++++
Smooth muscle
 PMwIFf++++++++++++
 IA4++++++++++++
Cardiac muscle
 Binucleatedg++++++++++++
 IA4 + MF-20++++++++++++
Adipocytes
 PMwRICVh++++++++++++++++++
 SBB++++++++++++++++++
 ORO++++++++++++++++++
Cartilage nodules
 Rd + PCMHi++++++++++++++++++
 C11C1++++++++++++++++++
 HC-II++++++++++++++++++
 D1-9++++++++++++++++++
 AB1.0++++++++++++++++++
 PF/AIcBI++++++++++++++++++
Bone nodules
 3D-Matrixj++++++++++++++++++
 WV1D1++++++++++++++++++
 MP111++++++++++++++++++
 vK++++++++++++++++++
Fibroblasts
 SpM + Pk++++++++++++
 1B10++++++++++++
Endothelial cells
 Cobblestonel++++++++++++
 P1H12++++++++++++
 P2B1++++++++++++
 P8B1++++++++++++
 P2H3++++++++++++
Table 2. CD marker expression profilea
 Prenatal donorsPostnatal donors
CM-SkM1CF-SkM1Mean ± SDNHDF1NHDF2PAL3PAL2Mean ± SD
  • a

    CD Marker expression detected by immuno-flow cytometry. Results are expressed as percentage of cells exhibiting positive staining for cell surface CD markers from a gated population of 10,000 cells.

CD31.501.401.451.360.190.00.00.42
CD40.050.550.300.2611.360.260.02.97
CD80.590.760.680.380.710.201.600.72
CD11c0.431.200.820.240.890.240.00.34
CD330.820.710.760.200.110.200.00.13
CD340.011.800.9025.2073.2034.3018.8038.88
CD361.351.541.440.365.600.360.01.58
CD380.860.800.830.2616.120.260.04.16
CD450.050.740.400.306.210.320.432.42
CD9095.5087.0891.2994.9988.5994.9982.4090.24
CD1170.041.340.690.403.120.400.00.98
GlycoA1.181.311.240.2297.540.220.024.50
HLA-DRII0.050.740.400.360.990.360.00.43

When undergoing myogenesis, skeletal muscle recapitulates its developmental sequence. This sequence begins with single mononucleated stellate cells that stain intracellularly for myogenin (F5D, Developmental Studies Hybridoma Bank, DSHB; Wright et al., 1991), and then for sarcomeric myosin (MF-20, DSHB; Bader et al., 1982). Subsequently, bipolar cells that stain intracellularly for sarcomeric myosin appear. The bipolar cells then fuse to form multinucleated structures that stain intracellularly for sarcomeric myosin, fast-skeletal muscle myosin (MY-32, Sigma; Naumann and Pette, 1994), myosin heavy chain (ALD-58, DSHB; Shafiq et al., 1984), and human fast myosin (A4.74, DSHB; Webster et al., 1988). Skeletal structures undergoing myogenesis are further identified by their elongated structure, multiple nuclei, cross-striations, and spontaneous contractility (Pate et al., 1993; Rogers et al., 1995; Young, 2000; Young et al., 1992a, 1993, 1995, 1998a,1998b). Maximal expression of skeletal muscle markers was induced with 10−8 M dexamethasone within the first two weeks of culture.

A smooth muscle cell phenotype was identified by morphological criteria. This phenotype consisted of large polygonal cells containing intracellular stress fibers. The smooth muscle phenotype was defined in this study by immunocytochemical analysis using an antibody to smooth muscle alpha-actin (1A4, Sigma; Skalli et al., 1986). Maximal expression of smooth muscle alpha-actin was induced with 10−8 M dexamethasone within the first 2 weeks of culture.

A cardiac myocyte phenotype was identified by morphological criteria. This phenotype consisted of binucleated cells exhibiting centrally located nuclei, and cross-striations. The cardiac myocyte phenotype was defined in this study by immunochemical analysis using colabeling of antibodies for both smooth muscle alpha-actin (IA4) and sarcomeric myosin (MF-20) (Eisenberg et al., 1997; Eisenberg and Markwald, 1997). Maximal expression of smooth muscle alpha-actin colocalizing with sarcomeric myosin in binucleated cells was induced by treatment with 10−8 M dexamethasone within the first 2 weeks of culture.

Adipogenic cells were identified by morphologic criteria. They appeared as polygonal cells containing multiple intracellular refractile vesicles. Adipocytes were verified by the presence of intracellular vesicles containing saturated neutral lipid by means of histochemical staining with Sudan Black-B (Chroma-Gesellschaft, Roboz Surgical Co, Washington, DC; Young et al., 1993) and Oil Red-O (Sigma; Humason, 1972). Maximal expression of adipogenic markers was induced by treatment with 10−9 M dexamethasone within the first 2 weeks of culture.

Chondrogenic structures develop along a progressional sequence. Initially, stellate cells express intracellular staining for type-II collagen [collagen pro type-II (C11C1, DSHB; Holmdahl et al., 1986; Johnstone et al., 1998) and human-specific collagen type-II (II-4CII, ICN Biomedicals, Aurora, OH; Burgeson and Hollister, 1979; Kumagai et al., 1994)]. Intracellular staining for type-IX collagen (D1-9, DSHB; Ye et al., 1991) follows. The cells become circular, expressing refractile pericellular matrix halos, and aggregate into nodules. The pericellular matrix halos demonstrate extracellular staining for type-II collagen and type-IX collagen and histochemical staining for sulfated glycosaminoglycans using Alcian Blue at pH 1.0 (Chroma-Gesellschaft; Young et al., 1993, 1998a,1998b) and Perfix/Alcec Blue (Fisher Scientific Co., Norcross, GA/Aldrich Chemical Co., Milwaukee, WI). Loss of staining after preincubation with chondroitinase-AC and keratinase (Young, 2000; Young et al., 1993, 1998a,1998b, 1999) confirms the presence of glycosaminoglycans containing chondroitin sulfate and keratan sulfate within the pericellular matrix halos stained with Alcian Blue pH 1.0. Maximal expression of intracellular chondrogenic markers was induced by treatment with 10−7 M dexamethasone by the end of 2 weeks of culture. Cartilage nodule formation with extracellular staining characteristic of chondrogenic cells occurred by 4 weeks when cultures were treated with 10−7 M dexamethasone.

Putative osteogenic structures develop along a progressional sequence. Initially, stellate cells express intracellular staining for bone sialoprotein (WV1D1, DSHB; Kasugai et al., 1992) and osteopontin (MP111, DSHB; Gorski et al., 1990). The stellate cells then form circular cells that continue to exhibit intracellular staining for bone sialoprotein and osteopontin. The round cells aggregate to form a three-dimensional extracellular matrix. The extracellular matrix stains with antibodies to bone sialoprotein and osteopontin. It also exhibits histochemical staining for calcium phosphate using the von Kossa procedure (Silber Protein, Chroma-Gesellschaft; Young, 2000; Young et al., 1993, 1995, 1998a,1998b, 1999). The von Kossa procedure stains for divalent cations. The identity of calcium as the cation stained (rather than zinc or magnesium) was confirmed using preincubation with EGTA (a chelating agent specific for calcium ions) rather than EDTA (a chelating agent for divalent cations) (Young, 2000; Young et al., 1993, 1995, 1998a,1998b, 1999). Maximal expression of intracellular osteogenic markers was induced with 10−9 M dexamethasone by the end of 4 weeks of culture. Mineralized nodule formation with extracellular staining specific for osteogenic cells occurred by 6 weeks when cultures were incubated with 10−9 M dexamethasone. Fibroblasts were identified by their spindle-shaped or polygonal-shaped morphology. The fibrogenic phenotype was verified by immunocytochemical staining with antibodies to human fibroblast surface protein (1B10, Sigma; Ronnov-Jessen et al., 1992). Maximal expression of the fibrogenic marker was induced by treatment with 10−8 M dexamethasone after 2 weeks of culture.

Putative endothelial cells develop along a progressional sequence. Initially, stellate cells express intracellular staining for human-specific endothelial cell surface marker (P1H12, Accurate, Westbury, NY; Solovey et al., 1997), peripheral endothelial cell adhesion molecule, PECAM (P2B1, DSHB), vascular cell adhesion molecule, VCAM (P8B1, DSHB; Dittel et al., 1993), and E-selectin (P2H3, DSHB). The stellate cells then form cobblestone-shaped cells, which are present individually or in sheets. Maximal expression of endothelial markers was induced by treatment with 10−8 M dexamethasone for 2 weeks of culture.

Secondary antibodies consisted of biotinylated anti-sheep IgG (Vector), biotinylated anti-mouse IgG (Vector), or were contained within the Vecstatin ABC Kit (Vector). The tertiary probe consisted of avidin-HRP contained within the Vecstatin ABC Kit (Vector). The insoluble HRP substrates VIP Substrate Kit for Peroxidase (blue, Vector), DAB Substrate for Peroxidase (black, Vector), and AEC Staining Kit (red, Sigma) were used to visualize antibody binding. The different substrates were utilized to allow for multiple sequential staining of the same culture wells.

Flow Cytometry

Aliquots of CM-SkM1, CF-SkM1, NHDF1, NHDF2, PAL3, and PAL2 cells at less than Hayflick's limit were thawed and seeded at 105 cells/1% gelatinized T-75 flasks in PM-B, and allowed to expand at 37°C in a 95% air/5% CO2 humidified environment. After expansion, cells were released with trypsin and resuspended in medium. The cells were then centrifuged and resuspended in wash buffer (Dulbecco's phosphate buffered saline without Ca+2, Mg+2 [Cellgro, MediaTech] supplemented with 1% FBS [HyClone] and 1% (w/v) sodium azide, NaN3 [Sigma]) at a concentration of 1 × 106 cells/ml. Cell viability was greater than 95% by the Trypan blue dye [GIBCO] exclusion technique (Young et al., 1991, 1993) and greater than 98% by the propidium iodide [Calbiochem-Novabiochem Corporation, La Jolla, CA] exclusion technique (Sasaki et al., 1987). One hundred microliters of cell preparation (1 × 105 cells) were stained with saturating concentrations of fluorescein isothiocyanate- (FITC), phycoerythrin- (PE), allophycocyanin (APC), or perdinin chlorophyll protein- (PerCP) conjugated CD3, CD4, CD8, CD11c, CD33, CD34, CD36, CD38, CD45, CD90, CD117, glycophorin-A, and HLA-II (DR), or isotype matched controls (Becton Dickinson, Inc. San Jose, CA). Briefly, cells were incubated in the dark for 30 min at 4°C. After incubation, cells were washed three times with wash buffer and resuspended in 0.5 ml of wash buffer. Flow cytometry was performed on a FACSort™ (Becton Dickinson) flow cytometer. Light scatter (Fig. 2) identified cells. Logarithmic fluorescence was evaluated (4 decade, 1,024 channel scale) on 10,000 gated events. Analysis was performed using Cellquest™ software (Becton Dickinson). The presence or absence of staining was determined by comparison with the appropriate isotype control. Gated events were scored for the presence of staining for a CD marker if more than 25% of the staining was above its isotype control. Percentages of cells per 10,000 gated events are shown in Table 3. A mean value above 10%-gated cells is considered positive for any given CD marker. Each cell line was run in triplicate, for a sample size of n = 3. Statistical analysis was performed on the pooled data from the six cell lines.

Table 3. CD cell surface markers on hematopoietic and neuronal cells
 CD3CD4CD8CD11cCD33CD36CD38CD45CD117Gly-A DR-II
  • a

    Kishimoto et al., 1997.

  • b

    Mizguchi et al., 1995; Rohn et al., 1996.

T-cellsa+++++
Monocytes/macrophagesa++++++
Natural killer cellsa++++
Granulocytesa+++++
Myeloid progenitor cellsa++++
Erythroid cellsa+
Some neuronal cellsb++

Statistical Analysis

Flow cytometry was performed in triplicate for each cell line. CD34 and CD90 were divided into two groups and the data pooled: (a) those derived from prenatal human tissues (n = 6) and (b) those derived from postnatal human tissues (n = 12). The two groups were analyzed by one-way analysis of variance, using the ABSTAT computer program (Anderson-Bell Corp., Arvada, CO).

RESULTS

Stem Cell Identification

The identities of the putative reserve stem cells present within male and female human fetal, adult, and geriatric cell populations were examined using an insulin/dexamethasone bioassay. Small numbers of phenotypic alterations in morphological appearance (1–5% expressed phenotypic markers) consistent with skeletal muscle myotubes, adipocytes, cartilage nodules, and bone nodules (data not shown) were produced upon treatment with insulin in all six cell populations at 30 cell doublings (Table 1). Larger numbers of phenotypic alterations were produced by treatment with dexamethasone. The morphologically altered cells stained for phenotypic expression markers indicating the presence of muscle cells, adipocytes, cartilage, bone, fibroblasts, and endothelial cells (data not shown, but similar to Fig. 1A–Z). These morphological alterations occurred in all six human stem cell populations at 30 cell doublings (Table 1). Above 70 cell doublings, insulin had no effect on the cells (Table 2), whereas dexamethasone altered both the phenotypic expression markers (6–10% expressed phenotypic markers) and morphological appearance of the cells (Table 2). Dexamethasone induced phenotypic expression markers for muscle, fat, cartilage, bone, fibroblasts, and endothelial cells. Muscle (skeletal, smooth, and cardiac) phenotypes were identified by antibody staining and morphological appearance. For example, the skeletal muscle phenotype was identified by intracellular staining for myogenic markers (Figs. 1A–E). These cells eventually formed multinucleated linear structures (Fig. 1F). Smooth muscle cells were identified by their polygonal shape and by the presence of stained intracellular stress fibers (Fig. 1G). Early cardiac myocytes were identified by a binucleated appearance and combined intracellular staining (Fig. 1H). Cells of the adipogenic (fat cell) lineage were identified as cells with intracellular refractile vesicles (Fig. 1I) containing saturated neutral lipids (Figs. 1J,K). Cells of the chondrogenic lineage were initially recognized by their intracellular staining for chondrogenic markers (Figs. 1L–N). These cells formed aggregations of rounded cells (Fig. 1O), which eventually formed discrete chondrogenic nodules (Fig. 1P,Q). Cells of the osteogenic lineage were initially recognized first by intracellular (Fig. 1R,S) and then extracellular localization of bone markers. These cells then aggregated and formed an overlying three-dimensional extracellular matrix (Fig. 1T,U). Cells of the fibroblastic lineage were initially recognized by their spindle- or polygonal shape and staining for human fibroblastic marker (Fig. 1V). Cells of the endothelial lineage were recognized by staining for endothelial markers (Figs. 1W–Z).

Figure 1.

NHDF2 cells at 80 cell doublings (A–I, K–O, R–T, V–Z) or PAL3 cells at 150 cell doublings (J, P, Q, U) treated for 14 (B, G, I), 28 (O), 42 (J, P, Q, T, U), or 56 (A, C, D–F, H, K–N, R, S, V–Z) days in medium containing 1%-HS9 (A–I, K–O, R–T, V–Z), 1%-HS3 (J, Q, U), or 10%-HS3 (P) and 10−6 M (A–P, R–T, V–Z) or 10−10 M (Q, U) dexamethasone. Cells were photographed at 25× (O), 40× (T, Y), 50× (Q, U), 100× (A–E, G, H, J–N, P–S, U–Z), or 125× (I) magnification in either brightfield (A–E, G, H, J–N, P–S, U–Z) or phase contrast (F, I, O, T) microscopy. A: Mononucleated cells staining intracellularly for myogenin (F5D). B: Mononucleated cells staining intracellularly for sarcomeric myosin (MF-20). C: Trinucleated cell staining intracellularly for skeletal muscle fast myosin (MY-32). D: Mononucleated cells staining intracellularly for myosin heavy chain (ALD58). E: Mononucleated cells staining intracellularly for myosin fast chain (A4.74). F: Two linear structures (arrows) containing multiple nuclei. G: Mononucleated cells staining intracellularly for smooth muscle alpha-actin (IA4). H: Binucleated cell co-staining intracellularly for sarcomeric myosin (MF-20) and smooth muscle alpha-actin (IA4). I: Mononucleated cells with intracellular refractile vesicles. J: Mononucleated cells with intracellular vesicles stained for saturated neutral lipids (Sudan Black-B). K: Mononucleated cells with intracellular vesicles stained for saturated neutral lipids (Oil Red-O). L: Mononucleated cell staining intracellularly for type-II collagen (CIIC1). M: Mononucleated cells staining intracellularly for type-II collagen (II-4CII). N: Mononucleated cells staining intracellularly for type-IX collagen (D1-9). O: Aggregation of cells with pericellular matrix halos. P: Nodule stained for chondroitin sulfate and keratan sulfate glycosaminoglycans (AB pH 1.0). Q: Nodule stained for sulfated moieties (Perfix/Alcec Blue). R: Mononucleated cells staining intracellularly for bone sialoprotein II (WV1D1). S: Mononucleated cells staining intracellularly for osteopontin (MP111). T: Three-dimensional matrix (arrow) overlying cell cluster. U: Nodules stained for calcium phosphate (von Kossa). V: Mononucleated cells stained for fibroblast specific protein (1B10). W: Mononucleated cells stained for human endothelial cell surface marker (P1H12). X: Mononucleated cells stained for peripheral cell adhesion molecule (P2B1). Y: Mononucleated cells stained for vascular cell adhesion molecule (P8B1). Z: Mononucleated cells stained for E-selectin (P2H3).

The data suggest that both progenitor cells (insulin-accelerated morphologies) and pluripotent cells (dexamethasone-induced morphologies) were present in the populations of putative human stem cells at 30 cell doublings (less than Hayflick's limit).

Flow Cytometric Analysis

The six cell populations were screened by immunochemistry coupled with flow cytometry for the presence of CD3, CD4, CD8, CD11c, CD33, CD34, CD36, CD38, CD45, CD90, CD117, glycophorin-A, and HLA-II (DR) (Table 3). Ten thousand cells per cell line were examined in gated populations designated as R1 (Fig. 2A). Greater than 98% of those cells were viable at time of cell screening as assessed by propidium iodide staining (Fig. 2B). Postnatal cells from adult and geriatric donors exhibited positive staining for CD34 (Fig. 3A). Staining for CD34 was not detected in the isolated prenatal cells (data not shown). All populations examined exhibited positive staining for CD90 (Fig. 3B). Postnatal cells from adult and geriatric donors expressed staining for both CD34 and CD90 staining (Fig. 3C). By contrast, the fetal populations only expressed CD90 (Table 3). No cells positive for CD34 but negative for CD90 were found in any population examined. On average, staining was negative for CD3, CD4, CD8, CD11c, CD33, CD36, CD38, CD45, CD117, and HLA-II (DR) (Table 3) in all populations examined. Glycophorin-A, CD38, and CD4 were positive only in the NHDF2 group (Table 3).

Figure 2.

A: Flow cytometry of FSC (forward scatter) × SSC (side scatter) showing R1 gated cell population of NHDF1 used for analysis. A similar R1 gate was used to analyze CM-SkM1, CF-SkM1, NHDF2, PAL2, and PAL3. B: Flow cytometry of forward scatter × PI (propidium iodide) of total cell population of CM-SkM1 used for analysis. The R4 gate shows cells within the total population that were propidium iodide negative, indicating greater than 98.9% viability of the cells. Similar plots for CF-SkM1, NHDF1, NHDF2, PAL2, and PAL demonstrate a range of cell viability from 98.4–99.5%.

Figure 3.

Flow cytometry of cluster differentiation markers. Horizontal axis denotes forward scatter (0 to 1,000 linear scale) and vertical axis denotes side scatter (0 to 1,000 linear scale). NHDF1 propagated to 30 cell doublings after harvest and analyzed with antibodies to cell surface cluster differentiation markers. A: CD34 vs. CD45. B: CD90 vs. CD45. C: CD34 vs. CD90.

Statistical analysis was performed for the presence of CD34 and CD90 in prenatal versus postnatal donors. For CD34 the percentage of cells labeled per 10,000 cells was as follows. Results are expressed as percentage of cell ± SEM, with the number of samples in parenthesis (n = #). Prenatal donors: CM-SkM1, 0.0100 ± 0.10 (n = 3); CF-SkM1, 1.8000 ± 0.502 (n = 3); mean, 0.9050 ± 0.459 (n = 6). Postnatal donors: NHDF1, 25.20 ± 0.528 (n = 3); NHDF2, 73.20 ± 0.972 (n = 3); PAL2, 18.80 ± 0.771 (n = 3); PAL3, 34.30 ± 4.28 (n = 3); mean, 38.88 ± 6.44 (n = 12). For CD34, analysis of variance demonstrated that the postnatal samples were significantly different from the prenatal samples: F(1,35) = 15.94, P < 0.003.

For CD90, the percentage of cells labeled per 10,000 was as follows: prenatal donors, CM-SkM1, 95.50 ± 1.81 (n = 3); CF-SkM1, 87.08 ± 0.886 (n = 3); mean, 91.29 ± 2.09 (n = 6). Postnatal donors: NHDF1, 94.99 ± 0.413 (n = 3); NHDF2, 88.59 ± 1.000 (n = 3); PAL2, 82.40 ± 0.095 (n = 3); PAL3, 94.99 ± 2.347 (n = 3); mean, 90.24 ± 1.46 (n = 12). For CD90, analysis of variance demonstrated that the postnatal samples were not significantly different from the prenatal samples: F(1,35) = 5427, P = 0.

DISCUSSION

Positive Staining for CD Markers in Human Mesenchymal Stem Cells

The functional significance of the expression of the cell surface cluster differentiation markers CD34 and CD90 expressed by the human stem cells used in this study is uncertain. CD34 is known to be expressed on committed and uncommitted hematopoietic precursor cells, small vessel endothelial cells and on some cells in nervous tissue (Lin et al., 1995; Ratajczak et al., 1998; Suda et al., 2000; Watt and Chan, 2000). Experiments using a cDNA clone have characterized CD34 as a sialomucin (Simmons et al., 1992; Suda et al, 2000). CD34 is thought to be involved in the regulation of the differentiation of hematopoietic stem cells, with some suggestion that it is a cell adhesion molecule (Gallacher et al., 2000; Lin et al, 1995; Nakamura et al., 1999). Clinicians have extensively utilized monoclonal antibodies to CD34 to purify hematopoietic stem cells and progenitor cells for use in autologous bone marrow transplantation (Hogan et al., 1997; Siena et al., 2000). In addition, selection for cells expressing CD34 may be employed to isolate cells in clinical applications for hematopoietic gene therapy (Bunnell et al., 1999; Sutherland et al., 1993).

CD90, also known as Thy-1, is expressed on hematopoietic cells (Imbert et al., 1998; Murray et al., 1999a; Stewart et al., 2000), thymic nurse cells (Bodey et al., 1999), neuronal tissue (Barlow and Huntley, 2000), renal mesangial cells (Tamura et al., 2000), placenta (Bukovsky et al., 1999) and some connective tissues (Saalbach et al., 1999). Human peripheral blood cells positive for both CD90 and CD34 include hematopoietic stem cells capable of producing multiple hematopoietic lineages (Murray et al., 1999a,b; Stewart et al., 2000). A function has not yet been assigned to CD90 (Saalbach et al., 1999), but it may play a role in signal transduction in T lymphocytes, as it is linked to pathways involving tyrosine phosphorylation (Lancki et al., 1995). The protein is considered to be part of the immunoglobulin superfamily because it shares some homology with immunoglobulins. Interestingly, because Thy-1 is expressed on brain tissue as well as T lymphocytes, this protein may play a role in the development of ataxia-telangiectasia. This disorder is characterized by lesions in both neurologic and immunologic function (Gatti, 1991; Teplitz, 1978).

The putative adult and geriatric stem cell populations expressed both CD34 and CD90 on the cell surface (as analyzed by flow cytometry), whereas the putative fetal stem cells expressed CD90 alone (Table 3). This finding may be important. It is possible that the cells positive for either CD34 or CD90 observed in the putative stem cell populations used in this study are derived from neuronal or connective tissue progenitor cells that survived in culture. The stem cell populations used for flow cytometry were at less than Hayflick's limit. Programmed cell senescence occurs after Hayflick's Limit (50–70 cell doublings) has been achieved (Hayflick, 1963, 1965). Because the stem cell populations used in this study had replicated fewer times than Hayflick's limit, they could theoretically still contain progenitor and differentiated cells. When cells from all populations were incubated with insulin, a known accelerator of phenotypic expression (Young, 2000; Young et al., 1992a, 1993, 1998a, 1998b), they did not demonstrate phenotypic markers for connective tissue cells (fibroblasts) (neuronal markers were not assayed in this study). This result suggested that progenitor cells for the connective tissue lineage were not present within our isolated cell populations. In addition, the cells positive for both CD34 and CD90 are unlikely to be derived from cells of neuronal or connective tissues, as cells from these tissues are not known to coexpress CD34 and CD90. The only previously described cell population positive for both CD34 and CD90 belongs to the hematopoietic stem cell lineage. We do not believe that the cells studied in this paper belong solely to the hematopoietic lineage, because of their ability to differentiate and then express phenotypic markers from multiple mesodermal lineages (muscle, adipocytes, cartilage, bone, fibroblasts, and endothelial cells) (Fig. 1). Rather, we believe that we have found another population that expresses both CD34 and CD90, in addition to the hematopoietic stem cell population previously described. The full significance of the presence of both CD34 and CD90 on these cells remains to be understood.

Negative Staining for CD Markers in Human Mesenchymal Stem Cells

In contrast to the findings for CD34 and CD90, 11 antigens were found to be absent on the cell surface of the cells examined (Table 3). These markers were CD3, CD4, CD8, CD11c, CD33, CD36, CD38, CD45, CD117, glycophorin-A, and HLA-II (DR). The significance of these findings is unknown at this time. These particular cell surface CD antigens, however, have been ascribed only to differentiated cells within the hematopoietic (and neuronal) lineages, i.e., T-cells, monocytes/macrophages, natural killer cells, granulocytes, myeloid progenitor cells, erythroid cells, and some neuronal cells.

The absence of these eleven surface markers characteristic of differentiated hematopoietic cells on the fetal, adult, and geriatric cells used in this study has at least two possible explanations. First, the putative stem cells examined may lack the capability under normal circumstances to differentiate along hematopoietic lineages. If this hypothesis is correct, these markers may never appear on differentiated lineages of these cells. Second, if these putative stem cells have the capability to differentiate along hematopoietic lines, the absence of the eleven differentiation markers may indirectly indicate that the cells studied are more primitive stem cells.

Our current research is aimed at understanding the putative stem cells located within connective tissue matrices. To that end we have isolated these cells from fetal, adult, and geriatric human donors of both genders and have begun characterizing their cell surface cluster differentiation antigens. Our initial study (Young et al., 1999) reported that human reserve stem cells displayed CD10, CD13, CD56, and MHC Class-I markers. We now report the expression of CD90 and varying amounts of CD34 in human reserve stem cells. We suggest that cell surface CD markers such as CD10, CD13, CD34, CD56, CD90, and MHC Class-I could be used in conjunction with flow cytometry and cell sorting as a second step in isolating more purified populations of these cells after an initial cell harvest.

Acknowledgements

The authors would like to express their thanks to John Knight for photographic assistance. The following antibodies: F5D developed by W.E. Wright, MF-20 developed by D.A. Fischman, ALD-58 developed by D.A. Fischman, A4.74 developed by H.M. Blau, CIIC1 developed by R. Holmdahl and K. Rubin, D1-9 developed by X.-J. Ye and K. Terato, WV1D1 developed by M. Solursh and A. Frazen, MP111 developed by M. Solursh and A. Frazen, P2B1 developed by E.A. Wayner and G. Vercellotti, P8B1 developed by E.A. Wayner and G. Vercellotti, and P2H3 developed by E.A. Wayner and G. Vercellotti were obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biological Sciences, Iowa City, IA.

Ancillary